IRSST - Institut de recherche Robert-Sauvé en santé et en sécurité du travail

Development of a Voice-based Method for Objective Occlusion Effect Measurement


When hearing protectors or hearing aids are worn, the obstruction of the ear canal can cause an undesirable and unpleasant effect, called the “occlusion effect,” which is manifested by a feeling that one’s own voice sounds different, with low frequencies amplified or distorted. The occlusion effect is often a source of discomfort that may lead the users of hearing protectors or hearing aids to misuse or to simply remove them. For workers, removal of hearing protectors can expose them to potentially dangerous noise levels, thus losing the benefit of wearing them. It is therefore important to quantify this effect in order to better understand it or to guide users in the choice of the most effective protector. The occlusion effect is characterized by an increase in low frequency sound pressure level in the occluded ear canal, induced by structure-borne excitation (speech, chewing, breathing, etc.). Thus, on the face of it, the measurement of the occlusion effect could be done “simply” by measuring this increase in noise level when the ear is occluded. Unfortunately, although measurement methods exist to quantify the occlusion effect in human subjects, there is no consensus in the scientific community as to the best method to use, and none of these are standardized.

One promising approach is to measure the occlusion effect using microphones placed in the ear canal and one’s own voice as the source of excitation. This project aims to (1) study the robustness of such a method for objective measurement of the occlusion effect, (2) propose an original alternative for assessing the occlusion effect, based on a single measurement, and (3) propose appropriate performance indicators. Trials have been conducted with two types of excitation in human subjects to analyze how these types of excitation affect the occlusion effect. Thus, various voice excitations (enumeration of random numbers, holding of the vowels /i/ and /ə/ at different levels of effort were used, as well as bone conduction excitation (bone vibrator and mastication).

The detailed analysis of the results led to the following findings: (i) it is possible to define a robust single value indicator that reflects the occlusion effect; (ii) vocal effort has little effect on the measured objective occlusion effect; (iii) use of random numbers rather than vowels (/i/ and /ə/) facilitates the application of the vocal effort method and reduces inter-subject variability; (iv) as expected, voice excitation produces a lower occlusion effect than bone conduction excitation (bone vibrator or mastication); (v) mastication as the source of excitation produces a similar average occlusion effect to that obtained with a bone vibrator; (vi) bone vibrator excitation produces higher inter-subject variability than voice excitation with random numbers.

The voice-based method presented in this report is relatively easy to implement and requires little difficult-to-operate equipment. In that sense, it is well suited to potential field use and could pave the way for developing a standardized method. In the long term, it should facilitate the choice of protectors by making it possible to obtain performance indicators related to the occlusion effect, in addition to the already existing performance indicators related to the sound attenuation provided by protectors.

Additional Information

Category: Research Report
Research Project: 2017-0032
Online since: March 30, 2021
Format: Text